Search papers, labs, and topics across Lattice.
Affiliation:
1
0
3
A verification layer powered by LLMs can filter out unsafe robot actions, achieving 97% containment of adversarial attacks while maintaining high decision-making precision.
